Estimação de estados e parâmetros de uma junta soldada

• Ano de defesa: 2023
• Autor(a) principal: Paviotti, Daniel Scopel
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Espírito Santo; BR; Mestrado em Engenharia Mecânica; Centro Tecnológico; UFES; Programa de Pós-Graduação em Engenharia Mecânica
• Palavras-chave em Português: Problema inverso; Filtro de partículas; Aproximação de erros; Soldagem; subject.br-rjbn; Engenharia Mecânica
• Link de acesso: http://repositorio.ufes.br/handle/10/16859

The welding processes, widely used in industry, use a large amount of heat to make the union of parts by melting materials. Immediately after the melting process, the cooling begins, so studying the temperature field is very important to understand the material's behavior. The present work shows the efficiency of the particle filter named Sampling Importance Resampling filter (SIR) when solving an inverse problem for estimation of the temperature of the weld bead after the process, as well as it is possible to minimize the error through the Approximation Error Model (AEM). Using a simplified model and comparing the results with the direct model was possible to observe the performance of the particle filter through the Root Mean Square (RMS) and the Relative Error (REL) of the monitored measurements. The computational cost is also a data to be observed showing that, as the number of particles increases, the time to compute the results also changes. Another essential parameter monitored was the torch speed once it directly influences the temperature field. The results found with the filter were satisfying and are according to the available literature. The application of AEM to improve the results also is satisfying when compared with the results without the application of the method. This method applied with the chosen filter is not typical for this kind of problem which can open a way for new studies on the proposed theme.